Building a machine-learning model to sharpen planning lead times

National industrial products distributor. Planning lead times didn't reflect real supplier performance across a large import and domestic supplier network — inflating safety stock and inventory across the distribution centres.

• Assessed the inventory benefit achievable with improved lead-time accuracy
• Built a cost-effective machine-learning prediction model for planning lead times — analysing the import and domestic supplier network, DCs, product categories and inbound supply
• Implemented model processes to refresh planning lead times on a semi-regular basis

• More accurate planning lead times feeding replenishment
• Quantified inventory reduction benefit case
• Repeatable, low-cost model process owned by the client